Wednesday, 28 April 2010

Farewell X Search, Hello One Search

The X Search service will no longer be available from the Library home page as of this Thursday evening, 29th April. The ability to search across multiple databases is now provided by One Search.

Perfect for New Researchers
We know how daunting it can be to start searching for academic literature - even choosing which link to follow from the Library & Computing web site is a challenge if you are not familiar with the array of resources we provide for you. One Search allows you to do one search across them all. And it's as simple as Google.

And Great for Experienced Ones
Studies show that students and faculty prefer the quality information found in libraries, but find searching on the Web simpler and more straightforward. One Search is designed to mimic open Web search methods, delivering the quality content that only the library holds. Now, you can simply enter a search term in the search box and the service will return - nearly instantaneously - a list of the physical and digital materials from the library's many collections that are relevant to that search. Users click-through to articles or find the book on the shelf in the library.

Try it:

Get the code for this widget

One Search's technical advantages over X Search
One Search is a web scale discovery tool, like Google it creates it's own index (in our case from Tropicat, ResearchOnline@JCU and our ejournal subscriptions) from published content. And it's fast.

X Search on the other hand is a 'federated search engine' which means it takes your query and translates it into a query for each of the databases (which each index publish content their own way) you are searching, waits for the results from all those databases, which have variable ranking alogorithms, response times and availability, merges and deduplicates them before presenting them to you - which can take an inordinate amount of time. It does not search Tropicat or ResearchOnline@JCU

One Search ingests structured metada into its purpose built index so complex searching can be performed across items from many different sources and it has a sophisticated ranking algorithm.

X Search takes the top 10 ranked records from each database in it's initial display of records which can lead to irrelevant records being listed in the top of the results. Because of the different capabilities of database interfaces there is a loss of more complex search capabilities because X Search has to fit the lowest common denominator.

X Search's search translators need constant updating as database interfaces change which can and has taken months of correspondence between X Search engineers and individual database providers.

One Search removes the vagaries of individual databases and providers from the process and is unaffected by changes and downtime from those sources.

When you use One Search to search the Library's collections you are only presented with items that are in the Library or available online fulltext. With X Search you are often presented with items that are only available via inter library loan. In One Search you can expand your search to include those items by ticking the 'Add results beyond your library's collection', as well as those open content items identified by other universities as worthwhile.

One Search is built on a rapid development model and comes with an API, which opens doors for us to develop new services as you as researchers express a need. For example we can build a tool that will search all the articles in a selected journal, even though that our holdings may be spread over several different platforms, or we could import Reserve Online items into One Search and then have a one click link from your LearnJCU course site to a listing of that course's readings.

No comments: